Methods of determining oxidizable constituents in gas mixtures are known. The conventional method is to measure a voltage between a working electrode and a reference electrode, which then permits an inference regarding the concentration of the gas constituent to be determined. Moreover, in the field of oxygen sensors for example, also known as lambda probes, it is known that the limit current induced between a reference electrode and a working electrode by transporting oxygen anions through a solid electrolyte body can be measured, permitting a determination of the equilibrium partial pressure concentration of oxygen according to the Nernst equation. The gases (O.sub.2, NO) to be measured pass through a diffusion barrier and are pumped out by electrochemical forces at the working electrode. However, there is not yet any satisfactory method of determining individual combustible gas constituents in gas mixtures with sufficient accuracy. In the case of ammonia in particular, there are no reliable systems at the present. Cross sensitivity to various gas constituents also has not been minimized satisfactorily. Another problem so far has been the temperature control of such sensors, which has previously required various complicated measurement arrangements.